Cell cycle regulation by long non-coding RNAs

The mammalian cell cycle is precisely controlled by cyclin-dependent kinases (CDKs) and related pathways such as the RB and p53 pathways. Recent research on long non-coding RNAs (lncRNAs) indicates that many lncRNAs are involved in the regulation of critical cell cycle regulators such as the cyclins, CDKs, CDK inhibitors, pRB, and p53. These lncRNAs act as epigenetic regulators, transcription factor regulators, post-transcription regulators, and protein scaffolds. These cell cycle-regulated lncRNAs mainly control cellular levels of cell cycle regulators via various mechanisms, and may provide diversity and reliability to the general cell cycle. Interestingly, several lncRNAs are induced by DNA damage and participate in cell cycle arrest or induction of apoptosis as DNA damage responses. Therefore, deregulations of these cell cycle regulatory lncRNAs may be involved in tumorigenesis, and they are novel candidate molecular targets for cancer therapy and diagnosis.


3
specific e3 ligases and eliminated via the ubiquitin-proteasome pathway [11][12][13]. The level of cell cycle regulators is precisely controlled by not only post-translational but also translational mechanisms. For example, several micro-RNAs (miRNAs) participate in cell cycle regulation through translational regulation [14]. MiRNAs are small non-coding RNA molecules containing 22 nucleotides, and negatively regulate translation through binding of the untranslated region of its target mRNAs [15]. The let-7 miRNA family negatively regulates cyclins A and D, and CDK4/6 and CDC25A [16]. The miR-15 family also inhibits the translation of cyclin D, CDK4, and CDC27 [17,18]. Interestingly, these let-7 and miR-15 family members may be involved in tumorigenesis since they are downregulated in various human cancers [16][17][18]. Alternatively, cyclin D1 is a target for not only let-7 and miR-15 miRNAs but also miR-19a, 26a, and 34a [15]. Furthermore, p27 Kip1 is targeted for regulation by the miR-181 family [19] and the miR-221 family [20]. The roles of other miRNAs in the expression of cell cycle regulators have also been reported [15]. Thus, it has been shown that the cell cycle regulators are critically and precisely controlled by e3 ligases and miRNAs both post-translationally and at the translational level.
Here, we focus on long non-coding RNAs (lncRNAs) involved in the regulation of the cell cycle through their various functions as epigenetic regulators, transcription factor regulators, post-transcription regulators and protein scaffolds [21,22]. LncRNAs are non-protein coding transcripts longer than 200 nucleotides, and can be divided into at least five categories based on their structural characteristics, including intergenic lncRNAs (lincRNAs), intronic lncRNAs, natural antisense transcripts, pseudogenes, and retrotransposons [23]. Recent mass-scale transcriptome analysis has revealed that many kinds of lncRNAs are transcribed in large amounts in the eukaryotic genome [24]. However, the question remains as to whether these lncR-NAs are merely by-products of the transcriptional units or have a critical function for biological processes. However, it has become clear that some of these lncRNAs participate in various biological processes such as genome imprinting, X-inactivation, development, differentiation, and cell cycle regulation [22,[24][25][26]. For example, HOTAIR, a wellinvestigated lncRNA, is involved in correct development The mammalian cell cycle is controlled by cyclin-dependent kinases (CDKs) and their related pathways. CDKs are activated via binding to their selected cyclins in specific phases of the cell cycle, following which they phosphorylate their target proteins. The CDK inhibitors (CKIs) negatively regulate the activities of CDKs and control the cell cycle. pRB regulates G1/S progression. The p53 pathway plays a role in DNA damage response as a gatekeeper of the genome. Several lncRNAs control the expression of cyclins-CDKs, CKIs, pRB and p53, and participate in cell cycle regulation. Some of these lncRNAs are induced by DNA damage and inhibit cell cycle progression by regulating these cell cycle regulators and tumorigenesis through recruiting the polycomb group (PcG) complex to its targeted HOX genes for their repression [26,27]. The PcG complex contributes to the epigenetic regulation of its target genes by forming Polycomb repressive complex 1 (PRC1) and 2 (PRC2). PRC2 participates in histone H3K27 methylation and, following histone H2AK119 monoubiquitination by PRC1, collaboratively represses target gene transcription. In addition to HOTAIR, several lncRNAs such as XIST, AIR, and KCNQ1OT1 also recruit chromatin modifiers including PcG and H3K9 methyltransferase G9a to their target loci [25,[28][29][30]. Moreover, ANRIL (antisense non-coding RNA in the INK4 locus) directly binds to PcGs and recruits them to the INK4 locus to promote gene silencing [31,32]. Thus, HOTAIR, XIST, AIR, KCNQ1OT1, and ANRIL function as epigenetic regulators by negatively regulating target gene transcription through recruitment of chromatin modifiers. Recently, several lncRNAs that participate in the expression of several cell cycle regulators have been reported (summarized in Fig. 1; Table 1). In this review, we introduce these lncR-NAs and discuss their functions in cell cycle regulation.

LncRNAs regulating cyclins and CDKs
Cyclins and CDKs are key players in cell cycle regulation (Fig. 1). NcRNA CCND1 , also called pncRNA (promoter-associated non-coding RNA), is transcribed from the upstream region of the cyclin D1 gene, CCND1, and negatively regulates cyclin D1. NcRNA CCND1 functions as a transcription factor regulator [33]. It is induced in a DNA damagedependent manner, and associates with and recruits TLS (translocated in liposarcoma) [34], an RNA binding protein. The ncRNA CCND1 -TLS complex is recruited to the CCND1 promoter to inhibit the activity of the coactivator, CBP/p300, thereby preventing CCND1 transcription (Fig. 2a). Thus, suppression of cyclin D1 regulated by the ncRNA CCND1 -TLS complex may participate in G1 arrest in response to DNA damage.
Gadd7 is an lncRNA involved in regulating CDK6 expression [35] in a posttranscriptional manner. TDP-43 (TAR DNA binding protein) binds to the 3′ untranslated region of CDK6 mRNA to stabilize it. Gadd7 is transcriptionally induced via DNA damage mediated by Uv and cisplatin [35,36], and binds to TDP-43 and dissociates from CDK6 mRNA. The CDK6 mRNA is then degraded, resulting in inhibition of the G1/S transition (Fig. 2b). Therefore, gadd7 negatively controls CDK6 expression, functioning as a translation regulator. Interestingly, gadd7 specifically controls mRNA stability for CDK6, but not CDK4, CDK2, or CCND1, by trapping TDP-43. The physiological relevance of the selective suppression of CDK6 remains to be determined. Gadd7 may be involved in the G1 checkpoint by collaborating with the lncRNA, ncRNA CCND1 , to downregulate the cyclin D1-CDK6 complex, thereby arresting cell cycle progression in response to DNA damage (Fig. 2a,  b). This may represent a novel G1-checkpoint cascade, but further studies are required.
MALAT1, an mRNA splicing mediator [37], is upregulated in several human cancers and contributes to cancer cell proliferation [38]. MALAT1 depletion results in arrest at G1 and promotes expression of p53 as well as p16, p21, and p27 in human fibroblasts [39] (Table 1). In contrast, MALAT1 depletion suppresses various genes involved in H19 lncRNA e2F1-mediated Downregulation of RB mRNA via miR675 (G1) [69][70][71] lncRNA-RoR p53-mediated Suppression of p53 mRNA translation (G2/M) [74] p53-induced eRNA p53-mediated Promotion of p53 target genes transcription (G1?) [75] loc285194 p53-mediated Growth inhibition by suppression of miR211 (G1?) [76] lncRNA-p21 p53-mediated Suppression of transcription of the target genes involved in apoptosis and cell cycle with hnRNA-K (G1?) suppression of β-catenin and Jun B mRNA translation [77,78] PANDA DNA damage Suppression of FAS and BIK transcription (G1?) [79] cell cycle progression such as the genes encoding cyclin A2 and Cdc25A, thereby arresting the cell cycle in G1. Moreover, in G2/M progression, MALAT1 is required for expression of B-Myb, which is involved in the expression of mitotic proteins such as cyclin B1, CDK1, FoxoM1, and PLK by controlling the splicing of B-Myb mRNA [39]. Therefore, MALAT1 may contribute to cell cycle progression in each phase by coordinated control of cell cycle regulators. Steroid receptor RNA activator (SRA) was identified as an lncRNA that binds to steroid receptors [40]. SRA forms the SRC-1 complex to activate transcription, mediated by steroid receptors such as progesterone receptor and estrogen receptor. It also binds to various other proteins such as myoD, and has multiple cellular functions such as myogenesis. SRA also binds to PPARγ and coactivates gene expression mediated by PPARγ. As such, SRA regulates adipogenesis and insulin sensitivity via PPARγ [41]. Additionally, SRA shows PPARγ-independent activity. Overexpression of SRA in pre-adipocytes downregulates the expression of cell cycle-promoting genes such as those encoding the cyclins [cyclins (A2, B1/2)], CDC20, MCMs (3,4,5,6), and CDT1. Conversely, these genes are upregulated by depletion of SRA. However, it remains to be elucidated whether SRA directly or indirectly suppresses the transcription of these genes, and further investigation into the mechanisms of SRA-regulated gene expression is required.

INK4 family inhibitors
The CDK inhibitory proteins, p16 ink4a and p15 ink4b (hereafter p16 and p15), bind to and inhibit CDK4 and 6, respectively, via their ankyrin repeats [3,42]. The p15 and p16 genes (CDKN2B and CDKN2A, respectively) are located at the INK4 locus together with the alternating reading frame gene, ARF [42]. ARF inhibits MDM2-dependent degradation of both p53 [43] and pRB [44]. Therefore, the expression of INK4 locus genes is critical for cell cycle regulation. The INK4 proteins are relatively stable, and their ubiquitin-dependent proteolysis is not particularly important for controlling their cellular levels. Therefore, the INK4 locus genes are mainly regulated by transcription. The participation of several transcription factors, including the eTS family [45], FOXO [46], and SP1 [47], has been reported. Moreover, the locus is regulated epigenetically. It has been suggested that PU.1 cooperates with  [48]. Moreover, PcG is recruited to the INK4 locus, thereby suppressing transcription via histone H3K27 methylation [49]. It has been suggested that antisense RNA transcribed near the p15 gene controls transcription of p15 [50]. Pasmant et al. identified an lncRNA, ANRIL, as an anti-sense transcript of the p15 gene in the INK4 locus [51]. Both our study and other research have revealed that ANRIL is involved in epigenetic repression of the transcription of the INK4 locus [31,32] (Table 1). we found that depletion of ANRIL by short hairpin RNA (shRNA) decreased the recruitment of SUZ12 to the INK4 locus and promoted the expression of p15 gene dramatically and p16 gene moderately, but had no effect on ARF [31]. SUZ12 is a component of the PRC2 complex. In contrast, Yap et al. [32] demonstrated that ANRIL binds to CBX7, a component of the PRC1 complex, in the chromatin fraction, and recruits PRC1 to the INK4 locus to mediate transcriptional suppression. Therefore, ANRIL binds to the PRC2 complex to recruit it to the INK4 locus, and then histone H3K27 methylation is mediated by eZH2 in the PRC2 complex. Next, PRC2 with ANRIL is recognized by CBX7, and the PRC1 complex is recruited to the region. Further, histone H2AK119 monoubiquitination is induced to repress transcription of the INK4 locus. Moreover, we demonstrated that depletion of ANRIL promotes growth arrest and induces senescence-associated beta-galactosidase in wI38 human fibroblasts [31]. Yap et al. [32] also suggested that CBX7-mediated suppression of the INK4 locus is involved in regulating cellular senescence. These reports strongly suggest that ANRIL participates not only in cell proliferation but also in suppressing premature senescence via the recruitment of PRC1 and PRC2 to the INK4 locus (Fig. 3a).
It is important to understand how ANRIL expression is regulated. we found that excess RAS signaling promoted by the introduction of activated H-RasG12v into wI38 fibroblasts suppressed ANRIL expression and induced p15 and p16, thereby arresting the cell cycle and inducing senescence-associated beta-galactosidase [31] (Fig. 3b). Recently, wan et al. [52] reported that ANRIL is induced by DNA-damaging agents via the ATM-e2F1 pathway, but p53 is not induced. Moreover, they suggested that depletion of ANRIL decreases homologous recombination after DNA double-strand breaks, although it is unclear whether ANRIL promotes DNA repair via the recruitment of PRC1 and PRC2 to the INK4 locus. Further studies are required on ANRIL function in response to cellular stresses. Moreover, Yang et al. found that lncRNA-HEIH is highly expressed in HBv-related hepatocellular carcinoma. It negatively regulates the expression of CDK inhibitors, such as p15, p16, p21, and p57, via interacting with eZH2, and then plays an important role in G0/G1 arrest [53] (Table 1). p18 ink4c (hereafter p18) is another INK4 family CDK inhibitor that also inhibits both CDK4 and 6 [3,54]. Recently, ink4c−/− mice have been shown to develop spontaneous pituitary adenomas [55], the frequency of which is enhanced by deletion of other CDK inhibitor genes [56]. The combined deletion of the p18 gene (CDKN2C) with the p16 gene is also found in human cancers [57]. Moreover, the expression levels of p16 and p18 are often inversely correlated during the progression of senescence [58]. It has been reported that transcription of the p18 gene is regulated by Menin-ReT-signaling and the PI3K-AKT pathway [59]. Du et al. [60]  lncRNA, HULC, negatively regulates the expression of p18 gene, which is located near the region containing HULC (Table 1). HULC was identified as an lncRNA upregulated in human hepatocellular carcinoma (HCC) [61] that is transcribed in a CReB-dependent manner [62]. Moreover, the expression of p18 is induced and suppressed by depletion and overexpression of HULC, respectively. The expression of p18 is inversely correlated with the expression of HULC in human HCC tissue specimens. Furthermore, the hepatitis B virus oncogene product, HBx, activates the HULC promoter via CReB to suppress the transcription of the p18 gene by upregulated HULC [60]. Downregulation of the p18 gene by HBx via HULC induction may contribute to the development of HCC, although it is unknown how HULC suppresses the transcription of the p18 gene.

Cip/Kip family inhibitors
The transcription of p57 Kip2 gene (CDKN1C), which is located at the KCNQ1 domain, is epigenetically suppressed as an imprinted gene on the paternal chromosome [63]. KCNQ1OT1 is paternally expressed as an antisense RNA of the KCNQ1 domain containing KCNQ1 and the p57 Kip2 gene [64]. KCNQ1OT1 functions as a recruiter that associates with the chromatin modifiers, PRC2 and G9a, and recruits them to the KCNQ1 domain to suppress the transcription of p57 Kip2 gene ( Table 1). As described above, lncRNA-HEIH downregulates the expression of not only the INK4 family inhibitors, p15 and p16, but also the Cip/ Kip family inhibitors, p21 and p57 [53].

LncRNAs regulating the pRB pathway
As described above, the tumor suppressor pRB is a critical regulator of G1/S progression [7,65]. It is well known that the expression of the RB gene is epigenetically silenced by methylation of the promoter in some cancers, including retinoblastoma [66]. Hypermethylation of the CTCF binding site in the RB promoter is mediated by the CTCF protein [67]. CTCF also regulates the expression balance between the IGF2/H19 locus together with DNA methylation of their promoters as an insulator of gene expression [68]. Interestingly, the H19 gene encodes a 2.9-kb lncRNA, and the H19 lncRNA is a precursor of miR-675 [69]. The expression of H19 lncRNA is mediated by e2F1 and promotes cell proliferation [70], but the mechanism is unknown. Tsang et al. [71] reported that the H19 lncRNA-derived miR-675 associates with the 3′ untranslated region of RB mRNA to negatively regulate pRB expression ( Fig. 4; Table 1). In human colorectal cancer, H19 lncRNA/miR-675 expression is inversely correlated with pRB expression [71]. Therefore, H19 lncRNA/miR-675 may be a critical negative regulator of the RB tumor suppressor pathway (Fig. 3) Fig. 4 Model showing the proposed mechanisms of lncRNA-mediated regulation of the RB pathway. pRB binds target transcription factors such as e2F and inhibits their activity in the G1 phase. Cyclin Ds-CDK4/6 phosphorylate pRB and activate e2F-mediated transcription in late G1, which regulates the expression of several growth-pro-moting genes and S phase entry. The transcription of H19 lncRNA from the H19 locus is mediated by e2F1. H19 lncRNA is processed to generate miR-675, which binds to RB mRNA and inhibits its translation pRB suppresses e2F-dependent transcription of H19 transcription via repression of the H19 promoter. Therefore, the H19-RB axis is self-regulated.
LncRNAs regulating the p53 pathway Another important tumor suppressor, p53, functions as the gatekeeper of the genome to control cell cycle arrest and apoptosis in response to DNA damage [65,72]. Although p53 is unstable, it is stabilized and activated via phosphorylation mediated by the ATM/ATR pathway in response to DNA damage. Moreover, p53 is also regulated via phosphorylation at various sites by specific kinases [73]. Zang et al. [74] reported that lncRNA-RoR negatively regulates p53 expression, thereby suppressing doxorubicin-induced G2/M arrest and apoptosis ( Table 1). Depletion of lncRNA-RoR leads to p53 accumulation, and overexpression of lncRNA-RoR suppresses p53 expression. LncRNA-RoR binds to phosphorylated heterogeneous nuclear ribonucleoprotein I (p-hnRNP-I) in cytoplasm and thereby suppresses p53 translation. The 28-base RoR sequence is sufficient for its function. Additionally, wild-type p53 binds to the RoR promoter to promote transcription of lncRNA-RoR, but mutant p53 does not bind to this promoter. This is a novel autoregulatory feedback loop that controls p53 levels (Fig. 5).
Recently, Melo et al. [75] reported that enhancer RNAs (eRNAs) are required for coordinated promotion between p53 target genes and p53-bound enhancer regions distant from the target gene, and participate in p53-dependent cell cycle arrest (Table 1). LncRNA loc285194 was suggested to have a tumor suppressor function, but its mechanism was unknown. Liu et al. found that loc285194 is induced by binding of p53 to its binding site in the promoter (Table 1). Moreover, they indicated that loc285194 binds to and inhibits miR-211, thereby downregulating miR-211-mediated cell proliferation [76]. Loc285194 is downregulated in human colon cancer specimens, and thus may contribute to the tumor suppressive function of p53 to inhibit miR-211 [76].
Huarte et al. [77] identified lncRNA-p21, which is transcribed near the p21 Cip1 gene (CDKN1A) as a p53-target gene. p53 directly binds to its binding element in the lncRNA-p21 promoter. Depletion of lncRNA-p21 alters  Fig. 5 Model showing the proposed mechanisms of lncRNA-mediated regulation of the p53 pathway. p53 controls cell cycle arrest, repair, and apoptosis in response to DNA damage. lncRNA-RoR binds to hnRNP-I and collaboratively suppresses p53 mRNA translation. This is an autoregulatory feedback loop that controls p53 levels. In response to DNA damage, p53 is stabilized and activated via phosphorylation mediated by the ATM/ATR pathway. p53 directly binds the target genes and regulates their expression to control cell cycle arrest, repair, and apoptosis. eRNAs are involved in promotion of p53-target genes in p53-dependent cell cycle arrest. p21 and lncRNA-p21, which is transcribed near the p21 Cip1 gene, are p53-target genes. lncRNA-p21 controls the expression of some p53-target genes. p53 function is partially mediated by gene regulation via lncRNA-p21 the expression of some p53-target genes except for p21 gene and inhibits apoptosis ( Fig. 5; Table 1). lncRNA-p21 binds to hnRNP-K and recruits it to the target genes, but the mechanism of target gene regulation is unknown. p53 function is partially mediated by gene regulation via lncRNA-p21-hnRNP-K. Moreover, Yoon et al. proposed that lncRNA-p21 functions as a modulator of translation. lncRNA-p21 associates with target mRNAs such as β-catenin and JunB in collaboration with Rck/p54 RNA helicase, and thus the translation of the target mRNAs is repressed [78]. Therefore, lncRNA-p21 regulates both transcription in the nucleus and translation in the cytoplasm.
PANDA (p21-associated ncRNA DNA damage-activated) was identified as a p21 promoter-derived transcript using ultra-high density tiling array of 56 cell-cycle genes. It is induced by DNA damage in a p53-dependent manner [79] (Table 1). PANDA binds to and inhibits NF-YA transcription factor, which limits the expression of proapoptotic genes such as FAS and BIK and results in the repression of apoptosis. PANDA is selectively induced in metastatic ductal carcinomas but not in normal breast tissue [79]. The results suggest that abnormal overexpression of PANDA may suppress apoptosis induced by DNA damage, which will accumulate and push the genome toward carcinogenesis.

Perspectives
Although the mechanisms of cell cycle regulation via cyclin-CDK, the p53/RB pathway, and the checkpoint pathway have been described in detail, recent studies on lncRNAs strongly suggest that lncRNAs control the expression of cell cycle regulators. Therefore, lncRNAs are critically involved in cell cycle regulation. However, it is unclear why lncRNAs might be deployed to regulate the cell cycle. As described in the "Introduction", lncRNAs involved in cell cycle regulation are classified into four groups. As shown in Table 1, ANRIL, lncRNA-HEIH, and KCNQ1OT1 are involved in epigenetic regulation of target gene transcription by collaborating with chromatin modifiers, which are classified as epigenetic regulators. ncR-NA CCND1 , SRA, PANDA, and lncRNA-p21 directly interact with the transcriptional machinery on the target genes and collaboratively regulate transcription as transcription factor regulators. Post-transcription regulators including gadd7, MALAT1, lncRNA-RoR, and loc285194 bind to their specific target mRNA to suppress translation and/or to modulate mRNA stability. SRA and MALAT1 also promote protein-protein interactions and are classified as protein scaffolds. Because the general cell cycle is closely associated with various cellular events as well as biological processes, it should be accurately regulated. Post-transcription regulators such as gadd7, MALAT1, H19 lncRNA, and loc285194 can rapidly and transiently suppress translation of their target genes. Transcription factor regulators such as ncRNA CCND1 , SRA, PANDA, and lncRNA-p21 that directly interact with the transcription machinery on the target genes may also participate in transient regulation. Alternatively, epigenetic regulators such as ANRIL, lncRNA-HEIH, and KCNQ1OT1 may have long-term effects on cellular senescence and imprinting because they mediate epigenetic regulation of cell cycle regulatory genes via chromatin modifiers. From this viewpoint, the cell cycle-regulated lncRNAs mainly control cellular levels of cell cycle regulators via various mechanisms, and may provide diversity and reliability to the general cell cycle.
It is interesting that many lncRNAs are associated with the DNA damage response. As shown in Table 1, 4 of 14 lncRNAs, lncRNA CCND1 , gadd7, ANRIL and PANDA, are induced by DNA damage. Another 4 lncRNAs, lncRNA-RoR, lncRNAp21, p53-induced eRNA, and loc285194, are induced in a p53-dependent manner, suggesting that they are induced by DNA damage. Therefore, these reported lncRNAs may participate in cell cycle arrest or induction of apoptosis as non-canonical DNA damage responses, whereas the ATM/ATR pathway is involved in a canonical DNA damage response to inactivate CDK activity as a DNA damage checkpoint. LncRNAs-mediated non-canonical pathways may ensure the response to DNA damage is diverse and reliable depending on the cellular context.
Considering the recent progress in lncRNA research, many lncRNAs that have a functional role in cell cycle regulation remain to be identified because the functions of only a small percentage of the total lncRNA population are understood. To clarify the roles of lncRNAs in cell cycle regulation, it should be determined how they regulate the target cell cycle regulators and which signaling pathways induce these lncRNAs. Since abrogation of the cell cycle is closely associated with cancer development and growth, cell cycle regulatory lncRNAs such as ANRIL and PANDA may have oncogenic properties. The importance of lncR-NAs in cell cycle regulation will be clarified by further pathological studies. Moreover, these cell cycle regulatory lncRNAs may be novel candidate molecular targets for cancer therapy or diagnosis.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.